many figures in backfile books could be upgraded to HTML5 figure elements

[eshellman]

this would improve accessibility. In many cases we could add semantic linkages from corresponing references in text.

for example, a book with this markup:

<div class="figcenter id003">
<img src="images/i_060.jpg" alt="" class="ig001">
<div class="ic001 caption">
<p>Fig. 5</p>
</div>
</div>

could be converted to

<figure id="figure_5" class="figcenter id003">
<img src="images/i_060.jpg" alt="" class="ig001">
<figcaption class="ic001 caption">
<p>Fig. 5</p>
</figcaption>
</figure>

by looking at classnames.

[eshellman]

This seems too difficult because changing the elements messes with the formatting. Instead, I will apply the appropriate aria roles and attributes. So instead:

<div class="figcenter id003" role="figure" aria-labelledby="caption_001">
<img src="images/i_060.jpg" alt="" class="ig001">
<div class="ic001 caption"  id="caption_001">
<p>Fig. 5</p>
</div>
</div>

note that this should not be done for new submissions - those should use figure and figcaption elements.